Run ML on AMD GPU

1. Windows version (>10, 1709)

2. Install minconda in C:\Users\amit\

What is conda? conda is package manager which will install python,other required packages.

3. Create python env using conda, Activate

//Create a new environment named ml_venv, install Python 3.11.5
cmd>
C:\Users\amit\source\repos\Python\ML>C:\Users\amit\miniconda3\Scripts\conda create --name ml_venv python=3.11.5
C:\Users\amit\source\repos\Python\ML>"c:\Users\amit\miniconda3\Scripts\activate" ml_venv
//Env is created here: C:\Users\amit\miniconda3\envs
(ml_venv) C:\Users\amit\source\repos\Python\ML>"c:\Users\amit\miniconda3\condabin\deactivate.bat"  //deactivate
(ml_venv) C:\Users\amit\source\repos\Python\ML>
            

4. Install the TensorFlow with DirectML package

(directml) C:\Users\amit\source\repos\Python\ML>pip install tensorflow-directml

5. Add 2 tensors on GPU

(directml) C:\Users\amit\source\repos\Python\ML>python
Python 3.6.13 |Anaconda, Inc.| (default, Mar 16 2021, 11:37:27) [MSC v.1916 64 bit (AMD64)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> import tensorflow.compat.v1 as tf
>>>
>>> tf.enable_eager_execution(tf.ConfigProto(log_device_placement=True))
>>>
>>> print(tf.add([1.0, 2.0], [3.0, 4.0]))
2023-10-26 10:50:50.890594: I tensorflow/stream_executor/platform/default/dso_loader.cc:97] Successfully opened dynamic library dxgi.dll
2023-10-26 10:50:50.900401: I tensorflow/stream_executor/platform/default/dso_loader.cc:97] Successfully opened dynamic library d3d12.dll
2023-10-26 10:50:51.365531: I tensorflow/core/common_runtime/dml/dml_device_cache.cc:250] DirectML device enumeration: found 1 compatible adapters.
2023-10-26 10:50:51.367876: I tensorflow/core/platform/cpu_feature_guard.cc:142] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2
2023-10-26 10:50:51.371564: I tensorflow/core/common_runtime/dml/dml_device_cache.cc:186] DirectML: creating device on adapter 0 (AMD Radeon R7 Graphics)